G-Square Architecture for Unified Analytics
G-Square Solutions provides a complete package/architecture for unified analytics, which helps its clients in achieving various analytics & machine learning tasks right from the ETL of data to its final analytical usage. It is a light weight and pocket friendly architecture, widely present in the market.
Please find the below architecture for broader elaboration:
The components of the architecture are:
ETL Toolkit – G-Square ETL toolkit is a highly sophisticated combination of various technologies such as Apache Airflow (latest technology in market used to author workflows), high level Python scripts & Data Connectors which are built in house and proprietary to G-Square Solutions. An alternative open source solution to this is Talend ETL toolkit.
Data Warehouse – This is the staging platform wherein the user stores the data and uses it for further analysis. PostgreSQL is used for storing structured data & Mongo DB is used to store documents/unstructured data. Complex level joins can be performed to collate data from both the databases. An alternative open source solution to this is Elastic Search.
Data Marts – Data Marts help the users or the departments to extract the data related to their interest which may also be used for querying and make analysis out of it. The data marts are created using MonetDB which is the best database for faster querying.
BI & Analytics – This is the layer where all the transformed data will be used for further analysis & AI related Task. Analysis is done by ‘Narrator Maker’ and the prepared analysis is sent in bulk to a group of branch level users with the help of ‘Narrator Viewer’ application. Integrate of the existing BI & reporting tool in this layer is possible. Usage of R & Python for various Machine Learning & data science activities is done.
Follow
